3 ms·

A decent case can be made for March 1 as the start of the year. With March 1 as the start of the year the extra day in leap years then comes on the last day of

by tzs 2y ago

A decent case can be made for March 1 as the start of the year. With March 1 as the start of the year the extra day in leap years then comes on the last day of the year. This makes dealing with leap years when calculating the day of the week from a date simpler. It also makes the names of September, October, November, and December make sense. Those names come from the Roman words for 7, 8, 9, and 10, which were the numbers of those months in the old Roman calendar which did start the year on March 1. The Romans switch the start to January sometime around 700 BCE giving us the current ridiculous situation where those months are two months later in the year than their names suggest.

kgwgk 2y ago
> The Romans switch the start to January sometime around 700 BCE That doesn’t look right.
- tzs 2y ago
  Well, that’s what the internet told me. But checking again it is also telling me 45 BC and 153 BC. The site that gave me the 700 BC date said it was done by king Numa Pompilius. Wikipedia’s article on him says that Plutarch is the source for the claim Numa moved the new year to January. But Plutarch was writing around the 2nd century and Wikipedia says his writings about Numa give unique information on the Roman calendar which suggests that there aren’t other sources. Probably not worth trusting a single source that old writing about something 900 years older.
wazoox 2y ago
The change was made by Caesar's reform of calendar in 46BC. January first was the date of consular elections, and Romans didn't used year number ab urbe condita ("since the foundation of Rome") regularly in daily life, instead they called the years by the running consuls' names. Therefore it made more sense for Romans to start the year at January first.
- kridsdale1 2y ago
  Sounds like the Japanese Emperor Era system.
athom 2y ago
Starting from March also sets up a neat pattern that can be used to generate accurate month/day values from day-of-year numbers, IF you number them from zero. You can also go back the other way. March through July follow a simple two-month cycle of 31 days, followed by 30, repeating at May and July. August breaks the cycle with 31 days, but it essentially restarts, and runs a second five months before it breaks again, at January (the 11th month, since we started from March). The third run obviously runs out after February, ending prematurely, but this doesn't really break the cycle, either. You just restart again with March! This allows you to write some (relatively) simple conversions using just a few divisors, along with some conversions from 1-based to 0-based numbering, and back. 153 days every five months, 61 every two, and of course 31 for a single month. Divide the remainder from each prior division by the next divisor, and the final remainder is your (0-based) day of the month. The month itself (taking 0 = March) is then the sum of each quotient multiplied by the associated count of months. Reversing the process, I'll leave to the reader.

by tzs 2y ago

kgwgk 2y ago

> The Romans switch the start to January sometime around 700 BCE That doesn’t look right.

tzs 2y ago
Well, that’s what the internet told me. But checking again it is also telling me 45 BC and 153 BC. The site that gave me the 700 BC date said it was done by king Numa Pompilius. Wikipedia’s article on him says that Plutarch is the source for the claim Numa moved the new year to January. But Plutarch was writing around the 2nd century and Wikipedia says his writings about Numa give unique information on the Roman calendar which suggests that there aren’t other sources. Probably not worth trusting a single source that old writing about something 900 years older.

wazoox 2y ago

The change was made by Caesar's reform of calendar in 46BC. January first was the date of consular elections, and Romans didn't used year number ab urbe condita ("since the foundation of Rome") regularly in daily life, instead they called the years by the running consuls' names. Therefore it made more sense for Romans to start the year at January first.

kridsdale1 2y ago
Sounds like the Japanese Emperor Era system.

athom 2y ago

Starting from March also sets up a neat pattern that can be used to generate accurate month/day values from day-of-year numbers, IF you number them from zero. You can also go back the other way. March through July follow a simple two-month cycle of 31 days, followed by 30, repeating at May and July. August breaks the cycle with 31 days, but it essentially restarts, and runs a second five months before it breaks again, at January (the 11th month, since we started from March). The third run obviously runs out after February, ending prematurely, but this doesn't really break the cycle, either. You just restart again with March! This allows you to write some (relatively) simple conversions using just a few divisors, along with some conversions from 1-based to 0-based numbering, and back. 153 days every five months, 61 every two, and of course 31 for a single month. Divide the remainder from each prior division by the next divisor, and the final remainder is your (0-based) day of the month. The month itself (taking 0 = March) is then the sum of each quotient multiplied by the associated count of months. Reversing the process, I'll leave to the reader.